1. Introduction


2. Data set

  • abstract
  • yearsActive
  • genre
  • recordLabel
  • instrument
  • occupation

2.1 Data Matrix


2.2 Abstract


3. Text Mining

3.1 Abstract

one-gram

one-gram token

two-grams

two-grams token


3.2 Token Matrix

  • tokens: 94 from 23070

4. Multi-value Classification

image from

4.1 Compress \(\mathcal{Y}\) information

Y-info.

Musicians

Instruments


4.2 PCA on Instrument Affinity Matrix

4.2.1 PC1 vs PC2

4.2.2 PC2 vs PC3

4.2.3 PC1 vs PC3


4.3 Correlation

Instrument

Musicians

5. Remark

  • The way out of multi-value classification
  • Group by PCA on instrument